I wanted to know about your opinion and maybe how did you solve this level.
Have you used libraries like tesseract-ocr, gocr, etc, or did you made an ocr logic yourself? Do you think it worth to make a simple ocr library from scratch, or this is reinvent of the wheel?